Noise ordinance and neighbour hearing — before a complaint arrives

Reading order

Name the sound object before you name a decibel. Normal cooling, generator test, emergency run, construction and deliveries are different sources at different hours. The thin noise page already asks for that matrix. This page adds the ordinance object and the neighbour object: which advice paper is being used, by which desk, and whether a PBL hearing has actually been sent.

Then name the desk. Boverket’s BFS 2020:2 and report 2020:8 sit with planläggning and building-permit review of new dwellings. Naturvårdsverket’s 2015 guidance sits with Environmental Code supervision and permitting of the noisy activity. Boverket’s types-of-noise page says general advice is not binding, but following it means the binding rule the advice belongs to has been met. Those are two letterheads. Do not collapse them into “the Swedish noise limit.”

Then name the neighbour paper. Chapter 9 sections 94 to 97 write when the building committee must hear neighbours on a lov or förhandsbesked. Chapter 2 section 9 writes that location and design must not cause danger to health or a significant inconvenience. A detaljplan samråd is a third paper. The complaints page is a fourth paper after operation. A hearing is not a consent, and consent for a measure closer than 4.5 metres to a boundary is a fifth paper on the same Boverket page.

• Sound object and operating case before a decibel.
• Boverket housing advice versus Naturvårdsverket activity guidance.
• A PBL neighbour hearing is not a neighbour consent.
• The complaints page already owns the supervision file after a complaint.

What an ordinance paper and a hearing paper actually do

Naturvårdsverket’s April 2015 guidance is a 22-page support document for Environmental Code supervision and permitting. The publication record says it replaces the earlier general advice RR 78:5 and the interim industrial-noise guidance. The PDF says the tabulated levels are a starting point for the individual inconvenience assessment at dwellings, schools, preschools and healthcare premises, outdoors at façade and outdoor seating. The day / evening / night starting points printed there are 50, 45 and 40 dBA as equivalent levels. The same pages say often-recurring impulses or clearly audible tones should normally lower the values by 5 dBA, and that night-time maximum levels above 55 dBA should not occur except on isolated occasions. Those are starting-point sentences. They are not a data-center statute. The complaints page already refused to reprint them as “the allowed Swedish decibel.” This page keeps that refusal and only names them as the activity-side starting points.

Boverket’s BFS 2020:2 and report 2020:8 measure a different desk. They apply when new dwellings are planned or given a building permit next to industrial and similar sound — fans, compressors and heat pumps are named examples. They are not written as the permit paper for a hall. Boverket’s types-of-noise page also says the advice does not cover shooting ranges, wind turbines, motor tracks, construction sites or temporary activities. Construction noise on a hall plot therefore stays in its own cell. General advice is not binding.

The neighbour-hearing sections in chapter 9 are a third mechanism. Section 94 writes the two cases that trigger a hearing. Section 95 writes the content of the notice. Section 96 writes who receives a sent notice and who is reached by the noticeboard. Section 94 also says a hearing is not required if it is obvious that a permit cannot be given, or that the application has no significance for the adjacent right-holder. A later change of the application needs a new hearing only if the change conflicts with the plan, except that those who already commented must always be heard again. Those are form rules. Boverket says skipping them is such a serious formal fault that the decision is likely to be quashed on review.

Chapter 2 section 9 remains the inconvenience test on the planning-and-building side. It is not a printed decibel. A neighbour’s local knowledge is, on Boverket’s page, one input to whether a measure is a significant inconvenience. It is not a vote that must be unanimous.

What named primary sources show

Boverket’s neighbour-hearing page is the dated form source. Keep sections 94 to 97, the sent-notice versus noticeboard split, the “not a consent” sentence and the later-appeal sentence attached to that page. Do not move them onto an Environmental Code complaint. The Planning and Building Act text in the register is the statute those sections sit in.

BFS 2020:2, report 2020:8 and the types-of-noise page are the housing-side ordinance sources. They name fans, compressors and heat pumps as examples and exclude construction sites. They do not publish a Swedish hall permit decibel. Naturvårdsverket’s 2015 PDF and its publication record are the activity-side starting-point sources. Keep 50 / 45 / 40 dBA labelled as starting points for an individual assessment, not as a national data-center limit.

The complaints page already owns chapter 26 supervision after a complaint. The noise-and-permit-timelines page already owns acoustic clocks on planning and environment. This page does not steal those matrices. Named Swedish halls stay on their own papers. No cited source here publishes a project-specific noise decision for Meta Luleå, Microsoft Staffanstorp, EcoDataCenter Falun or atNorth SWE04.

A matrix instead of a quiet slogan

Draw one page with paper rows and desk columns. BFS 2020:2 housing advice, Naturvårdsverket 2015 activity starting points, a PBL section 94 hearing, a detaljplan samråd and a later chapter 26 complaint are different rows. Empty cells stay empty. A 50 dBA starting point is a footnote on the activity row, not a hall permit.

Do not invent a Swedish data-center noise ordinance from Boverket’s dwelling advice or from Naturvårdsverket’s starting points. Those remain different desks from different years.

Common misreads

The first misread is to paste Boverket’s dwelling advice onto a hall permit as “the Swedish limit.” The advice is written for new housing next to industry. The second is to treat Naturvårdsverket’s 50 / 45 / 40 dBA starting points as a statute. They are starting points for an individual assessment. The third is to treat a neighbour hearing as a consent or a veto.

The fourth is to skip the hearing because the plot sits inside a detaljplan that already “allows industry.” Section 94 still triggers on a departure from the plan. The fifth is to treat a detaljplan samråd as the lov hearing. The sixth is to treat a later complaint as proof that no hearing was needed. The seventh is to invent a data-center noise chapter that neither Boverket nor Naturvårdsverket published.

What to ask next

Ask which operating case the acoustic paper actually models. Ask which desk is using which advice — BFS 2020:2 for nearby new housing, or the 2015 guidance for the activity. Ask whether a section 94 hearing has been sent, to whom, and on which date. Ask whether anyone has treated a comment as a consent.

Then open the thin noise page, the complaints page and the noise-and-permit-timelines page. If the file is silent about the hearing and the desk, record the gap. Do not fill it from a quiet-campus slogan.

1. Which advice paper is in play — BFS 2020:2 or NV 2015?
2. Has a section 94 notice actually been sent or posted?
3. Has a hearing been treated as a consent?
4. Has 50 dBA been reprinted as an allowed hall limit?
5. Has a data-center noise ordinance been invented?

What this does and does not prove

This page proves that Swedish primary sources already publish a PBL neighbour-hearing clock in chapter 9 sections 94 to 97, a Boverket housing-side industrial-noise advice paper, and a 2015 Naturvårdsverket activity-side starting-point paper, and that those papers are not a national data-center decibel.

It does not prove how loud a named Swedish hall is. It does not invent a permit month. It does not treat a hearing as a consent.

Empty cells stay empty. Until a later primary source names a Swedish hall’s acoustic decision with a dated operating case, the ordinance stack stays a method, not a census.